Breaking out of the Poverty Trap

Home , Abhijit Banerjee, Esther Duflo, James Foster (economist)

CHAPTER THREE Breaking Out of the Poverty Trap

Lindsay Coates and Scott MacMillan

Introduction In 2018, one of us visited a rural village in Bangladesh to speak to participants in a “graduation program,” a term used to describe programs designed to break the poverty trap with a boost of multiple, sequenced interventions. When we asked one woman what the program had changed for her, she brought out a piece of paper inviting her to a village event. Before she went through the program, her neighbors barely knew she existed, she said. Now she was a member of the community, invited to people’s homes and weddings. One hears echoes of this senti- ment from participants in similar graduation programs worldwide. Such stories illustrate just one of the many cruel aspects of ultra-poverty: those afflicted by it tend to be invisible—to neighbors, distant policymakers, and nearly everyone in between. The ultra-poor need to stop being invisible to policymakers. We need to pay closer attention to the poorest and the unique set of challenges they face, for without a better understanding of the lived reality of ultra-poverty, we will fail to live up to the promise of “leaving no one behind.” Without programs tailored for people in these circumstances, the extreme poverty rate will become increasingly hard to budge. We are already starting to see this reflected in global poverty data. For decades, the global extreme poverty rate, defined as the portion of humanity living below the equivalent of $1.90 per day, fell rapidly, from 36 percent in 1990 to 10 percent in 2015. Earlier in this decade, optimism took hold

With special thanks to Emily Coppel and Isabel Whisson.

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 41 9/6/19 1:57 PM 42 Lindsay Coates and Scott MacMillan

that we may even remove extreme poverty from the face of the earth.1 The World Bank has interpreted “eradication” as less than 3 percent, but optimism is now waning that we will reach even that milestone by 2030. The World Bank’s own 2018 Poverty and Prosperity report offers a stark warning: “To reach our goal of bringing extreme poverty below 3 percent by 2030, the world’s poorest countries must grow at a rate that far surpasses their historical experience.”2 Another report projects that climate change and forced displacement will cause another 100 million people to fall into extreme poverty by 2030.3 Even if 4 percent of the world’s population remains below the threshold in 2030, this will be an estimated 340 million people, more than the current population of the United States—hardly a footnote or a rounding error.4 This is unacceptable from a moral, rights-based standpoint. It is also an inefficient use of global human potential, creating less opportunity and progress for us all. Breaking the poverty trap so that hundreds of millions can become productive economic citizens is both an ethical impera- tive and sound economic policy. This chapter seeks to advance an understanding of the microeconomic and psychological reality of what it means to be ultra-poor, while pointing to an emerging set of scalable, science-based solutions that can break the trap. “The poor” are not a homogenous group, and even the term “extreme poor” is often used to lump together people facing very different circumstances. Using the graduation approach pioneered by BRAC as one example, this chapter will highlight ways to tackle ultra-poverty through the emerging “science of hope,” which posits that when coupled with skills and material support, an injection of well-founded hope and optimism into the lives of the ultra-poor can break the poverty trap. A growing body of evidence suggests that programs activating people’s sense of self-worth lead to improvements in employment, earnings, mental health, political awareness, and women’s influence in the household. There is also growing evidence to suggest that when it comes to ultra-poverty, purely economic boosts, including relatively quick fixes such as unconditional cash transfers or universal basic income, can fail to break the trap, while more holistic changes make the crucial difference.

1. Lowrey (2013). 2. World Bank (2018). 3. Hallegatte and others (2016). 4. Also concerning is that the majority of progress against extreme poverty in recent decades came from East Asia and the Pacific region, with much of it tied to China’s rise. This will not neces- sarily carry over into other regions. In sub-Saharan Africa, absolute numbers of the extreme poor are actually growing, from 278 million in 1990 to 413 million in 2015. See World Bank, 2018, “Poverty and Shared Prosperity 2018—Piecing Together the Poverty Puzzle.”

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 42 9/6/19 1:57 PM Breaking Out of the Poverty Trap 43

What Is Ultra-Poverty? Ultra-poverty is hard to define. The commonly accepted extreme poverty income threshold is currently $1.90 per day, and the ultra-poor tend to live on less than that. At levels that low, income- and consumption-based definitions stop being meaningful. Those who work closely with the ultra-poor observe that their suffering has a distinctive character, even compared to people just slightly better off. We define ultra-poverty as a sub-segment of extreme poverty characterized by material destitution and psychological despair so severe that mainstream development assistance and market-led solutions make no dent in it. It is among the clearest examples of a poverty trap: a self-reinforcing state of physical, material, and psychological deprivation seemingly immune to most interventions designed to boost people’s income and well-being. A useful lens through which to view ultra-poverty is economist Amartya Sen’s definition of poverty as a “deprivation of basic capabilities,” including the freedoms and choices that most of us take for granted, specifically “the substantive freedoms [a person] enjoys to lead the kind of life he or she has reason to value. In this perspective, poverty must be seen as the deprivation of basic capabilities rather than merely as lowness of incomes, which is the standard criterion of iden- tification of poverty.”5 Based on qualitative reports from development workers and researchers, a functional understanding of what ultra-poverty looks like has emerged, and it tracks closely with Sen’s definition.6 Far beyond simply having low incomes, the ultra-poor are bereft of any semblance of those substantive freedoms. Figures vary as to how many people fall into this category, as the data is spotty and the definition remains loose, but the number likely remains in the hundreds of millions. According to a 2007 report from International Food Policy Research Institute based on 2004 data, about 162 million people lived in ultra-poverty, which was defined at the time as living on less than 50 cents a day, in 1993 purchasing power parity (PPP) dollars, with an additional 323 million living in “medial poverty,” defined as living on between 50 and 75 cents a day.7 Accord- ing to a more recent estimate from 2017, 736 million people live in extreme poverty—that is, below $1.90 per day in 2011 PPP terms—and of these, more than half, or around 394 million people, are living in ultra-poverty.8

5. Sen (1999). 6. See “Impact Pathways: Stories from People Behind the Numbers,” from Akhter U. Ahmed and others, “The Impact of Asset Transfer on Livelihoods of the Ultra Poor in Bangladesh,” Research Monograph Series No. 39, BRAC Research and Evaluation Division, April 2009, (http:// ebrary.ifpri.org/utils/getfile/collection/p15738coll2/id/125248/filename/125249.pdf) 7. Ahmed and others (2007). 8. Reed and others (2017).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 43 9/6/19 1:57 PM 44 Lindsay Coates and Scott MacMillan

All poverty arises from a complex set of interrelated causes, often tied to fail- ures in politics and governance along with systematic marginalization of minori- ties, migrants, and—almost universally across cultures—women. For those at the very bottom, addressing the problem requires highly adapted approaches that account for the unique characteristics of ultra-poverty in each region or even community. The indicators of ultra-poverty are multi-modal: the Global State of Ultra-Poverty, which numbered the population of global ultra-poor at 394 million, defines ultra-poverty using the Multidimensional Poverty Index, a measure that uses methods developed by Sabina Alkire and James Foster of the Oxford Poverty and Human Development Initiative.9 The method assesses poverty across multiple indicators, including education, electricity, sanitation, and access to drinking water.10 It is challenging to promote common approaches to such complex, multifaceted problems, each with their own set of contextual realities. Worse yet, access to existing support programs—whether provided by government or civil society—is spotty across regions, which is again tied to discrimination as well as political and social marginalization. In some places, promoting access to a local government vaccination program might help solve a critical health need; in other places, such a program might not even exist for anyone. Despite the uneven progress and complexity of the diagnosis in any given context, interest in addressing ultra-poverty—and the capacity to do so—is growing among governments and global development institutions. The Partnership for Economic Inclusion (PEI), which is housed at the World Bank and advocates for accelerated innovation and scaling of the graduation approach,11 gives a snapshot of progress in its 2018 State of the Sector report.12 Governments from thirty-four countries are now engaged in graduation (sometimes also referred to as “productive inclusion”), double the number from 2016, reaching an estimated 14 million people.

The Power of Hope What creates a poverty trap? Development practice and a growing body of evidence suggest the ultra-poor may often be caught in a self-perpetuating trap of

9. Ibid. 10. Ibid. See Oxford Poverty and Human Development Initiative, “Global Multidimensional Poverty Index,” (https://ophi.org.uk/multidimensional-poverty-index/). 11. PEI is the successor to the Graduating the Poor Initiative, previously housed at Consultative Group to Assist the Poor (CGAP), which oversaw the first series of global graduation pilots to test the effectiveness of the model. 12. Arevolo, Kaffenberger, and de Montesquiou (2018).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 44 9/6/19 1:57 PM Breaking Out of the Poverty Trap 45

hopelessness, which sustains itself based on a person’s inability to imagine that a better world is possible.13 The trap of hopelessness likely arises in part from cognitive overload. The poor have many choices, and as unfashionable as it may sound, in many cases they have too many choices. They can walk several miles for their child’s booster immunization, or they can work that day. The choice is entirely theirs. The stress of being poor—including the uncertainty of paid labor, of not knowing where your next meal will come from, and the constantly looming threat of health shocks—is compounded by the necessity of making a multitude of important decisions like these every single day. Sendhil Mullainathan, an economist, and Eldar Shafir, a psychologist, explore the phenomenon of the “bandwidth tax,” which depletes one’s finite cognitive resources.14 The authors cite studies on parenting: Air traffic controllers, on a busy day at work, when they have to make hundreds of potentially life-or- death decisions, tend to be worse parents that evening, compared to less busy days; the children of people on food stamps behave worse at school at the end of the month, when the family’s food stamps are running out, suggesting that parents’ daily worries trickle down to their children, affecting their faculty for self-control.15 Mullainathan and Shafir call this the “present-day cognitive load of making ends meet,” and it is especially acute for those living in ultra-poverty. Some might object that it is overly paternalistic to ascribe people’s suffering to hopelessness, as though all that is needed is a shift in mindset. This would be a misunderstanding of the nature of both the problem and the solution. In fact, many of the freedoms we take for granted actually result from what Esther Duflo calls a “subtle paternalism.” For instance, people in richer countries usually do not have to worry about boiling their water before drinking it, because the state has decided it is better for them to have treated water coming from their faucets.16 Far from infringing on Sen’s “substantive freedoms,” this actually gives us more freedom to pursue other sources of satisfaction and meaning, for it removes a source of stress from our lives and allows us to apply our cognitive resources elsewhere. For most people struggling with ultra-poverty, even modest goals such as owning chickens are out of reach—and known to be out of reach—without a material boost.

13. Abhijit Banerjee and Esther Duflo cite evidence to suggest that, much like nutrition, psychological factors have physiological and material consequences. In many cases, when incomes rise for the very poor, food expenditure rises roughly proportionally to overall consumption. If the poverty trap were caused by undernutrition only, hungry people would have invested most of their new income in food, so they could work more and earn more money. See “Poor Economics,” Public Affairs, 2011, 22–28. 14. Mullainathan and Shafir (2013). 15. Ibid., 155–57. 16. Duflo (2012).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 45 9/6/19 1:57 PM 46 Lindsay Coates and Scott MacMillan

We are not suggesting that psychological factors like stress, hopelessness, or the bandwidth tax are the sole factors keeping people in ultra-poverty. Hope and confidence alone will not buy you chickens; in many cases, people’s pessi- mism is entirely warranted. But we have little doubt these psychological factors are at work and too often ignored. The question is whether a material boost alone would break the trap and thus shift the psychology, or whether psychology remains a barrier unto itself. We posit, based largely on qualitative observations of how people break the poverty trap, that in most cases of ultra-poverty, even if material conditions change through a sudden positive shock (such as a one-time income boost or transfer of assets), the psychological factors remain an obstacle. There is not yet overwhelming evidence for or against this hypothesis. It is one that can and needs to be empirically tested, as we will soon discuss.

Graduation: A Case Study in the Science of Hope The graduation approach has gained recognition over the past five years as one of the most rigorously tested interventions in international development. Broadly speaking, it is a sequenced set of services tailored to the culture, context, and population of the ultra-poor, designed to give a one-time boost to break the poverty trap. These interventions generally include livelihood training, transfers of cash and productive assets, and encouragement of savings, all facilitated through in-person coaching. They focus on the needs of the household as a whole, with women usually as the primary program participants. When adapting graduation globally, BRAC applies a framework approach to ensure that programs meet four goals: social protection, livelihood promotion, financial inclusion, and social empowerment. For details of the graduation methodology used by BRAC in Bangladesh, the largest and longest-running such program in the world, see box 3-1. The role of the coach is key, for this is how these programs build hope. Each participant is assigned a caseworker or coach tasked with conducting regular individual check-ins, typically at the participant’s home. In addition, the coach meets regularly with small groups of participants who live in the same village or neighborhood. In the group setting, participants learn from and receive encouragement from one another. These recurring touchpoints remind the participant that change is possible and support will be available along the way. The entire program is notably time-bound, usually running for about two years, during which a person living in dire poverty can be expected to transform his or her life, “graduating” onto the economic ladder—at least to the bottom rung of it— where he or she will enjoy a degree of control and independence, often for the first time in his or her life. Emphasis is on the participants to lead their own

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 46 9/6/19 1:57 PM Breaking Out of the Poverty Trap 47

Box 3-1. An Overview of BRAC’s Ultra-Poor Graduation Program BRAC’s Ultra-Poor Graduation (UPG) program (formerly called Targeting the Ultra-Poor) is the first and largest such program in the world. Launched in Bangladesh in 2002, the program has reached 1.9 million households, or more than 7 million people, as of 2019. The process of participant selection begins with a rigorous period of targeting, including community meetings to identify the poorest people in the area. Program staff follow up with door-to-door visits, using questionnaires to help verify who qualifies. Participants are trained in certain self-employment skills—how to take care of a cow or small shop, for instance—and in a one-time transfer, they are given a productive asset, such as livestock, agricultural inputs, or goods for small trade. In the program’s earlier phases, participants were given limited consumption support in the form of cash and food to offset the opportunity cost of not engaging in menial wage labor or begging while they learned new livelihoods. Recognizing the improving poverty environment in Bangladesh, this provision has since been replaced with a “savings match,” whereby BRAC matches money participants put aside, which helps build a habit of saving. Participants also receive financial training, including saving for the future, budgeting, and using credit to help sustain or grow their micro-enterprises. Program staff make sure participants’ children are going to school, have balanced diets, are not at risk of child marriage, and ensure the household has access to healthcare, clean water, and a latrine. The program also mobilizes the community to help oversee aspects of the program, provide support where needed, and integrate the ultra-poor into local life, in part by setting up village development committees, which include local elites. The graduation rate of participants—that is, the percentage that leaves a rigorously defined category of ultra-poverty based on a set of predefined indicators and takes up more sustainable livelihoods—usually exceeds 95 percent.

progress, with the knowledge that at the end of the two-year period, they will be on their own. If this progress stalls, the caseworker will be there to help. To be sure, the “graduated” ultra-poor are still poor, but they have secured stable incomes for their families and are finally in a position where their hard work actually gains traction. They have broken free from the poverty trap, because the program has addressed the multifaceted nature of that trap, whereby any one of a multitude of factors—including lack of skills, low assets and capital, ill health and poor nutrition, battered confidence and hopelessness—could have weighed them down and prevented them from breaking free. BRAC’s original graduation program emerged from the organization’s recognition of its own failure to reach the poorest in Bangladesh, where the organization was founded in 1972. By the late 1990s, BRAC had grown to become a well-established microfinance provider in Bangladesh, offering credit services and training throughout the country. Staff recognized that although many

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 47 9/6/19 1:57 PM 48 Lindsay Coates and Scott MacMillan

women—for the microfinance program was exclusive to women—were using small loans to grow and expand their businesses, a substantial population was so marginalized they failed to benefit from microfinance or any other mainstream development programs that BRAC offered. Research and adaptation are central to the BRAC approach. Launched in 2002, the Ultra-Poor Graduation (UPG) program evolved as it encountered unforeseen challenges, its staff tweaking it to account for the changing poverty context in Bangladesh. This was often in response to studies and evaluations conducted by BRAC’s independent research unit, which has published twenty- five working papers since 2004, each examining a cultural nuance, challenge, or component of the program in depth. Topics have included how to effectively find and target participants, how to leverage relationships and engage with village elders to facilitate social inclusion, and whether credit combined with grants can address ultra-poverty in a more sustainable manner than grants alone.17 By 2007, qualitative and quasi-experimental studies suggested the two-year program was having a profound and long-lasting impact. Scholars from the Lon- don School of Economics began a randomized controlled trial (RCT), covering about eight thousand women in just over fourteen hundred villages. A working paper released in 2013 found significant gains versus a control group, including a 38 percent increase in income and “large and sustained impacts on the occupational choices and economic lives of the eligible poor.”18 Most notably, participants’ gains had persisted after four years, or two years after they had stopped receiving direct support, suggesting the program had indeed broken the poverty trap. The program also saw gains in self-reported happiness, value of assets, a reduction of anxiety, and other non-economic indicators of improved well-being. Since then, several follow-up surveys using the same baseline have found that participants continue to gain ground, even seven years after entering the program.19 Additional follow-up studies are underway. Following early indications of the effectiveness of BRAC’s graduation program in Bangladesh, interest grew in knowing whether the BRAC results could be replicated in other geographic and cultural contexts. Starting in 2006, the

17. Noor and others, “Stories of Targeting: Process Documentation of Selecting the Ultra Poor for CFPR/TUP Programme,” CFPR-TUP Working Paper Series No. 1, BRAC Research and Eval- uation Division, April 2004 (https://research.brac.net/new/publications/documentation); Naomi Hossain and Imran Matin, “Engaging Elite Support for the Ultra-Poor,” Development in Practice, November 2007; Jinnat Ara and others, “Walking on Two Legs: Credit Plus Grant Approach to Poverty Reduction,” CFPR-TUP Working Paper Series No. 25, BRAC Research and Evaluation Division (http://research.brac.net/new/staff/ultrappor). 18. Bandiera and others (2013). 19. Bandiera and others, “Labor Markets and Poverty in Village Economies,” Quarterly Journal of Economics, vol. 132, no. 2 (May 2017), 811–70; Balboni and others, “Transforming the Economic Lives of the Ultra-Poor,” Growth Brief, International Growth Centre, December 2015.

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 48 9/6/19 1:57 PM Breaking Out of the Poverty Trap 49

Figure 3-1. BRAC’s 2007 Targeting the Ultra-Poor Graduation Program

Source: Credit: Ampliﬁ er Strategies

Consultative Group to Assist the Poor (CGAP) and Ford Foundation collab- orated to lead a set of pilots in eight countries, testing adaptations of BRAC’s approach (box 3-2). The process culminated in 2015, when a group of researchers published the combined results from RCTs covering more than ten thousand households in six of those countries, comparing participants’ results to that of a control group one year after graduating from the program. The results were impressive across the board: income and revenues were significantly higher in the treatment group in every country, and household consumption was significantly higher in all countries except one, Honduras.20 The question of whether graduation could be a globally adaptable solution to extreme poverty was answered with a resounding yes. Graduation began gaining prominence in development circles, sparking a wave of investment in close to a hundred programs worldwide since the findings were released.

Comparing Graduation to Unconditional Cash Transfers As the graduation approach has gained traction, another category of interventions, unconditional cash transfers (UCTs), has caught the attention of donors

20. Banerjee and others (2015).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 49 9/6/19 1:57 PM 50 Lindsay Coates and Scott MacMillan

Box 3-2. Evidence Shows the Graduation Approach to Be a Globally Adaptable Solution to Extreme Poverty Between 2006 and 2014, the Ford Foundation, Consultative Group to Assist the Poor, and several international and local partners tested and adapted the graduation approach in ten pilot programs in eight countries. Randomized controlled trials were conducted in six of the sites—in Ethiopia, Ghana, Honduras, India, Pakistan, and Peru. Each program ran for two years, with the evaluations assessing results at the end of the program and one year after its completion. Program adaptations in each country were designed to take into account local market forces, barriers to economic activity facing the ultra-poor in each context, existing social protection programming with which to link participants, program eligibility criteria, and graduation indicators specific to the community. The studies found that every group of economic outcomes improved significantly relative to the comparison group, both immediately after the program ended and a year later. Graduation households’ consumption increased 16.4 percent in Ethiopia, 13.6 percent in India, and 10.2 percent in Pakistan, with similar improvements observed in savings, asset holdings, and food security. Impacts also extended to psychosocial well-being. In Ethiopia, Ghana, Honduras, India, and Peru, improvements in self-reported happiness and stress remained positive a year after the end of the program. Women also became more likely to be politically active and reported having greater input in household financial decisions. Most notably, the results were largely consistent across multiple contexts and partners, signaling that ultra-poor households faced similar con- straints across diverse contexts, which graduation was largely effective at addressing.a

a. Abdul Latif Jameel Policy Action Lab. 2015. “Building Stable Livelihoods for the Ultra-Poor,” Policy Bulletin, September.

and development agencies by recording strong gains with equally robust (albeit shorter-term) randomized impact evaluations. These programs are establishing the idea of cash benchmarking, testing the hypothesis that everything delivered during the course of a development intervention—be it training, technology, assets, education, health, or social support—might be better off delivered to participants as cash. If the intervention does not outperform cash along desired impact indicators, it should be scrapped in favor of UCTs—so the thinking goes. Compared to graduation programs, UCT programs can be politically appealing for governments, since a UCT program typically makes less effort to identify participants, a cost-saving measure that also avoids hard choices about whom to include. Many UCT programs use simple targeting criteria, such as whether a potential participant has a thatched or a metal roof.21 This absence of robust

21. This was one of the methods initially used by GiveDirectly, one of the UCT pioneers. More recently, the organization began targeting entire villages using national poverty data, then enrolling all households in a single village, akin to a universal basic income plan.

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 50 9/6/19 1:57 PM Breaking Out of the Poverty Trap 51

targeting makes it more difficult to meaningfully compare UCT programs with highly targeted graduation programs, since the two approaches may have heterogeneous results for different groups, based on participants’ initial level of poverty or other baseline characteristics. Perhaps the most well-known UCT studies are the ones on the GiveDi- rectly program in Kenya.22 The initial results showed broad, positive changes from unconditional cash transfers after nine months, including increased consumption and savings, investments in livestock and durable assets (mainly metal roofs), and increases in revenue from agriculture and business activities. Other studies of unconditional or near-unconditional cash transfers have also yielded initial positive results. The Ugandan government’s Youth Opportunities Pro- gram in northern Uganda tested the effect of giving large, near-unconditional cash grants (participants were merely required to state how much of the grant they would invest in training) to thousands of unemployed youth, with grant sums worth twice their average annual income. Recipients invested the grants mainly in vocational skills training, and after four years, business assets were 57 percent higher than a control group, work hours 17 percent higher, and earnings 38 percent higher.23 Cash transfers have led to improvements in non-economic indicators in other contexts, such as early marriage and teenage pregnancy, as in a 2011 study comparing conditional and unconditional cash transfers in Malawi.24 Newer evidence suggests these gains may fade over time, however. A follow-up on the Malawi study showed that the short-term benefits of monthly cash transfers largely evaporated after two years.25 The three-year results from the Give- Directly Kenya program, released in 2018, have led to multiple interpretations, including one that suggests the gains from cash had dissipated at the three-year mark. Though the results are subject to debate, there is even some evidence that the transfers had negative spillover effects, in the sense of having actually caused harm to non-recipients in the same village. The authors concluded that “cash transfers result in sustained increases in assets,” but that “long-term impacts on

22. Johannes Haushofer and Jeremy Shapiro, “The Short-Term Impact of Unconditional Cash Transfers to the Poor: Experimental Evidence from Kenya,” Quarterly Journal of Economics, vol. 131, no. 4 (November 2016); Johannes Haushofer and Jeremy Shapiro, “Household Response to Income Changes: Evidence from an Unconditional Cash Transfer Program in Kenya,” Princeton Univer- sity, November 2013 (www.princeton.edu/~joha/publications/Haushofer_Shapiro_UCT_2013. pdf); Johannes Haushofer and Jeremy Shapiro, “The Long-Term Impact of Unconditional Cash Transfers: Experimental Evidence from Kenya,” January 2018 (http://jeremypshapiro.com/papers/ Haushofer_Shapiro_UCT2_2018-01-30_paper_only.pdf). 23. Blattman, Fiala, and Martinez (2014). 24. Baird and others, “Cash or Condition? Evidence from a Cash Transfer Experiment,” Quar- terly Journal of Economics, vol. 126, no. 4, November 2011, 1709–53. 25. Baird, McIntosh, and Özler (2016).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 51 9/6/19 1:57 PM 52 Lindsay Coates and Scott MacMillan

other dimensions, and potential spillover effects, remain to be substantiated by future work.”26 The nine-year impact of the above-cited Youth Opportunities Program in Uganda was even more lackluster. While researchers found the program had strong economic effects for at least the first four years, after nine years, the control group and treatment groups had converged, reducing the long-term impact to zero.27 Despite the attention both graduation and unconditional cash transfers have received, few studies have directly compared graduation and unconditional cash transfer programs in the same setting.28 There are currently less than a handful of current and ongoing studies that do compare cash and graduation, and while they have produced notable results, none has produced generalizable evidence of cost-effectiveness. 29 This is a salient gap in the development literature, for such studies would measure the relative cost-effectiveness of two of the most successful approaches to poverty reduction. It would also allow one to make bolder claims about which approaches work best for whom, based on heterogeneous baseline characteristics of a given population. While we are advocates for the graduation approach, we are open to empirical evidence that may show that, for certain populations, UCTs could be a more cost-effective solution. As the authors of the GiveDirectly RCT argued: “What is needed now, in our view, are studies that compare the effect of cash transfers to those of other interventions that have been shown to be effective in improving outcomes in developing countries. For instance, are UCTs more or less effective than ultra-poor graduation programs . . . and on what dimensions?”30

26. Haushofer and Shapiro (2018). 27. Blattman, Fiala, and Martinez (2014). 28. A 2016 meta-analysis attempted a cost-effectiveness comparison of forty-eight programs in three categories (UCTs, graduation, and generic livelihood programs). Based on available evidence, the paper called the graduation approach “the clearest path forward to reduce extreme poverty in a sustainable manner,” with the caveat that additional direct comparisons were needed. See Munshi Sulaiman and others, “Eliminating Extreme Poverty: Comparing the Cost-Effectiveness of Liveli- hood, Cash Transfer, and Graduation Approaches,” CGAP and IPA, December 2016. 29. A study published in December 2016 made a direct comparison of a BRAC graduation program in South Sudan and a similarly expensive unconditional cash-transfer program. The study found “evidence of positive consumption effects from both treatments, but a persistent wealth effect only from [graduation].” See Reajul Chowdhury and others, “Valuing Assets Provided to Low-In- come Households in South Sudan,” University of California, Riverside, December 2016 (https:// economics.ucr.edu/pacdev/pacdev-papers/valuing_assets_provided.pdf). 30. Haushofer and Shapiro (2018).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 52 9/6/19 1:57 PM Breaking Out of the Poverty Trap 53

Taking Graduation to Scale

Donors and development agencies need to reevaluate how they think about investing in and delivering complex, multidimensional implementations at scale. The graduation community has proven that complex programs for the ultra-poor can be adapted successfully. Leading implementers such as BRAC offer technical assistance to governments and others looking to implement such programs.31 While global partnerships such as the Partnership for Economic Inclusion housed at the World Bank encourage the uptake and scaling of graduation programs by client governments, additional high-level guidance would spur much- needed action. One model could be the set of norms and standards developed by the World Food Program, World Bank, and other actors, on how to develop and implement effective school-feeding programs.32 Without additional funding, we risk leaving the ultra-poor behind and falling short of the extreme poverty eradication goals. There is a tendency among donors to favor relatively simple, high-impact interventions, whose results are easier to replicate and scale. Bed net distribution, vaccination, and cash transfer programs (both conditional and unconditional) have gained popularity, partly for these reasons. But the complex nature of the poverty trap necessitates multifaceted programs. While graduation is sometimes seen as too costly to scale, the long- term cost-benefit analysis works in its favor. Most of the major studies on graduation include cost-benefit analyses showing gains that outlast the period of direct support, making it likely that a single upfront investment is a more effective way of helping the poor than many smaller payments over time.33 BRAC’s success in Bangladesh can largely be credited to the sustained funding it received for over a decade, which enabled years of iteration and adaptation to address the complexity of challenges facing the ultra-poor. This included large sums of unrestricted funding from the UK and Australian aid agencies as part of a Strategic Partnership Arrangement.34 The sustained financing mechanism provided BRAC with uniquely flexible capital, akin to what Acumen calls “patient capital.”35 BRAC’s success with the program also challenges the notion that civil

31. BRAC assisted many of the earlier graduation adaptations and, as of 2019, is offering technical assistance to ongoing non-BRAC graduation programs in Kenya, the Philippines, Egypt, Leso- tho, Rwanda, and Egypt. 32. Bundy and others (2009). 33. Banerjee and others (2015); Bandiera and others (2013). The Bangladesh RCT estimated a benefit/cost ratio of 3.2, assuming economic benefits persist for 20 years with a discount rate of 5 percent. 34. BRAC/DFID/AusAID Strategic Partnership Agreement (SPA)—Terms of Engagement, (2012). 35. See Acumen, “Patient Capital” (https://acumen.org/about/patient-capital/).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 53 9/6/19 1:57 PM 54 Lindsay Coates and Scott MacMillan

society organizations cannot be effective agents of scale. In fact, a 2019 report from the UK’s public foreign aid watchdog cited BRAC’s graduation program as one of the “best buys in development.”36 While such arrangements have gone out of favor with many donors, the success of graduation is a validation of the argument for sustained, flexible investments in organizations that are committed to learning and iterating. When governments can be convinced of the value in this, they often provide the best mechanism to reach large numbers of the ultra-poor. Many of the largest graduation-style programs (outside Bangladesh) are operated by governments. Many build on existing government services, improving the effectiveness and reach of investments governments are already making in cash transfers, financial services, livelihoods, or healthcare. Thus, governments may already have many parts in place, which means the development of graduation programs can consist of a coordination or deepening of existing measures, rather than having to start from scratch. Ongo- ing pilots in the Philippines, Rwanda, and Kenya link participants to existing social assistance programs such as job readiness training and subsidized health insurance. New funds for graduation may also be available through innovative financing mechanisms. Development impact bonds (DIBs) are well suited to highly measurable and evidence-based approaches like graduation. In 2018, a development impact bond, worth $5.28 million and funded by the U.S. and U.K. aid agencies along with private donors, enabled Village Enterprise to reach twelve thousand ultra-poor households in Kenya and Uganda. As long as participants make progress based on pre-agreed indicators as measured by an external evalu- ator, payment is unlocked and private funders receive back the capital they put in.37 Like the “patient capital” model, the implementer is judged by the results they achieve, not how closely they follow a method pre-agreed with the funder. This should create incentives for iteration and adaptation to contexts, which is key to successful graduation programming. Moreover, with DIBs, governments can be the “outcome payer,” increasing government ownership of interventions to support ultra-poor citizens while delegating implementation.

Conclusion Policymakers need to understand the complex needs and unique set of challenges faced by the poorest segments of the human population. If we are to live

36. Independent Commission for Aid Impact, “DFID’s Partnerships with Civil Society Orga- nizations: A Performance Review,” April 2019 (https://icai.independent.gov.uk/html-report/csos). 37. See Village Enterprise, “The Village Enterprise Development Impact Bond,” (https://villageenterprise.org/our-impact/development-impact-bond/).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 54 9/6/19 1:57 PM Breaking Out of the Poverty Trap 55

up to the promise of leaving no one behind—if we are to truly invest in human capital, such that humanity’s untapped resources of talent and ingenuity are harnessed for global growth and the common good—we must start by better understanding the special character of ultra-poverty and why it is qualitatively different from other forms of poverty. Part of the reason is nonmaterial factors such as stress, hopelessness, and in the words of Mullainathan and Shafir, the constant “present-day cognitive load of making ends meet,” which bears down like a weight too great to lift. Poverty is, at heart, a deprivation of one’s capacity to be fully human—to dream, to plan, to arrange one’s imagination to see a better future. The graduation approach offers a way out. The challenge now is to mobilize donors and the development community to provide resources to scale up such programs so we can finally eradicate extreme poverty from the face of the earth. Admittedly, in the words of BRAC’s founder, Sir Fazle Hasan Abed, “asking policymakers to invest in optimism and self- worth may sound like a vague, soft-hearted appeal.”38 It is anything but that, thanks to the emerging science of hope. When combined with the right amount of material support, an injection of hope and self-esteem can break the poverty trap for millions.

References Abdul Latif Jameel. 2015. “Building Stable Livelihoods for the Ultra-Poor.” Poverty Action Lab, Policy Bulletin. Abed, Fazle H. 2018. “Building Human Capital Means Investing in the Science of Hope,” Thomson Reuters Foundation, October 17 (http://news.trust.org/ item/20181017120852-ak1io/). Acumen. n.d. “Patient Capital.” (https://acumen.org/about/patient-capital/). Ahmed, Akhter U., and others. 2007. “The World’s Most Deprived: Characteristics and Causes of Extreme Poverty and Hunger.” 2020 Discussion Paper 43, International Food Policy Research Institute (www.ifpri.org/publication/worlds-most-deprived). Ahmed, Akhter U., and others. 2009. “The Impact of Asset Transfer on Livelihoods of the Ultra Poor in Bangladesh,” Research Monograph Series No. 39, BRAC Research and Evaluation Division (http://ebrary.ifpri.org/utils/getfile/collection/p15738coll2/ id/125248/filename/125249.pdf). Ara, Jinnat, and others. 2018. “Walking on Two Legs: Credit Plus Grant Approach to Poverty Reduction,” CFPR-TUP Working Paper Series No. 25, BRAC Research and Evaluation Division (http://research.brac.net/new/staff/ultrappor). Arevolo, Ines, Michelle Kaffenberger, and Aude de Montesquiou. 2018. 2018 State of the Sector—Synthesis Report. Partnership for Economic Inclusion, 26.

38. Fazle Hasan Abed, “Building Human Capital Means Investing in the Science of Hope,” Thom- son Reuters Foundation, October 17, 2018 (http://news.trust.org/item/20181017120852-ak1io/).

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 55 9/6/19 1:57 PM 56 Lindsay Coates and Scott MacMillan

Baird, Sarah, Craig McIntosh, and Berk Özler. 2016. “When the Money Runs Out: Evaluating the Longer-Term Impacts of Two-Year Cash Transfer Program.” University of California, San Diego (http://gps.ucsd.edu/_files/faculty/mcintosh/ mcintosh_research_SIHR.pdf). Balboni, Clare, and others. 2015. “Transforming the Economic Lives of the Ultra- Poor.” Growth Brief, FIRMS, International Growth Centre, 4. Bandiera, Oriana, and others. 2013. “Can Basic Entrepreneurship Transform the Economic Lives of the Poor?” Working Paper, International Growth Centre (www. theigc.org/wp-content/uploads/2014/09/Bandiera-Et-Al-2013-Working-Paper.pdf). ———. 2017. “Labor Markets and Poverty in Village Economies,” Quarterly Journal of Economics, vol. 132, no. 2 (May 2017), 811–70. Banerjee, Abhijit, and others. 2015. “A Multi-faceted Program Causes Lasting Progress for the Very Poor: Evidence from Six Countries,” Science, vol. 348, no. 6236,1260799-3. Banerjee, Abhijit, and Esther Duflo. 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty (Philadelphia: Public Affairs). Blattman, Christopher, Nathan Fiala, and Sebastian Martinez. 2014. “Generating Skilled Self-Employment in Developing Countries: Experimental Evidence from Uganda.” Quarterly Journal of Economics, vol. 129, no. 2, 697–752. Bundy, Donald, and others, “Rethinking School Feeding: Social Safety Nets, Child Development, and the Education Sector,” World Food Program and the World Bank, 2009. BRAC/DFID/AusAID Strategic Partnership Agreement (SPA)—Terms of Engagement. 2012. (https://dfat.gov.au/about-us/publications/Documents/brac-spa-terms- engagement.pdf). Chowdhury, Reajul, and others. 2016. “Valuing Assets Provided to Low-Income Households in South Sudan.” University of California, Riverside. (https:// economics.ucr.edu/pacdev/pacdev-papers/valuing_assets_provided.pdf). Duflo, Esther. 2012. “Human Values and the Design of the Fight Against Poverty.” Tanner Lectures (draft). (https://www.povertyactionlab.org/sites/default/files/ documents/TannerLectures_EstherDuflo_draft.pdf). The Economist. 2013. “Poverty: Not Always with Us” (www.economist.com/ briefing/2013/06/01/not-always-with-us). Hallegatte, Stephane, and others. 2016. “Shock Waves—Managing the Impacts of Climate Change on Poverty.” Climate Change and Development Series (Washington, D.C.: World Bank). Haushofer, Johannes, and Jeremy Shapiro. 2013. “Household Response to Income Changes: Evidence from an Unconditional Cash Transfer Program in Kenya.” Princeton University (www.poverty-action.org/sites/default/files/publications/ Haushofer_Shapiro_UCT_2013.pdf). ———. 2016. “The Short-Term Impact of Unconditional Cash Transfers to the Poor: Experimental Evidence from Kenya.” The Quarterly Journal of Economics, vol. 131, no. 4, 1973–2042. ———. 2018. “The Long-Term Impact of Unconditional Cash Transfers: Experimental Evidence from Kenya” (https://jeremypshapiro.appspot.com/papers/Haushofer_ Shapiro_UCT2_2018-01-30_paper_only.pdf). Hossain, Naomi, and Imran Matin. 2007. “Engaging Elite Support for the Ultra-Poor,” Development in Practice, vol. 17, no. 3 (June), 3.

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 56 9/6/19 1:57 PM Breaking Out of the Poverty Trap 57

Kristof, Nicholas, “The Power of Hope Is Real,” New York Times, May 21, 2015 (www. nytimes.com/2015/05/21/opinion/nicholas-kristof-the-power-of-hope-is-real.html). Lipton, Michael. 1986. “Seasonality and Ultra-Poverty.” IDS Bulletin 17.3 (https:// michaellipton.files.wordpress.com/2012/02/seasonalityandultrapoverty1986.pdf). Lowrey, Annie, “Is It Crazy to Think We Can Eradicate Poverty?” New York Times, May 5, 2013 (www.nytimes.com/2013/05/05/magazine/is-it- crazy- to- think- we- can- eradicate-poverty.html). Mullainathan, Sendhil, and Eldar Shafir. 2013. Scarcity: Why Having Too Little Means So Much (New York: Times Books). Noor, Marufia, and others. 2004. “Stories of Targeting: Process Documentation of Selecting the Ultra Poor for CFPR/TUP Programme,” CFPR-TUP Working Paper Series No. 1, BRAC Research and Evaluation Division (https://research.brac.net/ new/publications/documentation). Oxford Poverty and Human Development Initiative. n.d. “Global Multidimensional Poverty Index” (https://ophi.org.uk/multidimensional-poverty-index/). Sulaiman, Munshi, and Farzana Misha. 2016. “Comparative Cost-Benefit Analysis of Programs for the Ultra-Poor in Bangladesh.” Copenhagen Consensus Center (www. copenhagenconsensus.com/sites/default/files/sulaiman_misha_ultrapoor.pdf). Sulaiman, Munshi, and others. 2016. “Eliminating Extreme Poverty: Comparing the Cost-effectiveness of Livelihood, Cash Transfer, and Graduation Approaches.” Washington, D.C.: CGAP. Raza, Wameq A., and Ellen Van de Poel. 2016. “Impact and Spillover Effects of an Asset Transfer Program on Malnutrition: Evidence from a Randomized Control Trial in Bangladesh.” Working Paper. Reed, Larry, and others. 2017. The Global State of Ultra-Poverty 2017—A Global Analysis of the Greatest Opportunities to End Ultra-Poverty by 2030. Uplift and the RESULTS Educational Fund (www.ultra-poverty.org/assets/downloads/gsup-2017- report.pdf). Sen, Amartya. 1999. Development as Freedom (New York: Random House). Sturla, Kate, Neil Buddy Shah, and Jeff McManus. 2018. “The Great DIB-ate: Measurement for Development Impact Bonds.” Stanford Social Innovation Review (https://ssir.org/articles/entry/ the_great_dib_ate_measurement_for_development_impact_bonds). United Nations. 2013. A New Global Partnership—Eradicate Poverty and Transform Economies Through Sustainable Development (https://www.post2015hlp.org/ the-report/). Village Enterprise. n.d. “The Village Enterprise Development Impact Bond” (https:// villageenterprise.org/our-impact/development-impact-bond/). World Bank. 2012. World Development Report 2012: Gender Equality and Development (Washington, D.C.: World Bank), 5. ———. 2013. World Development Report 2013: Jobs. (Washington, D.C.: World Bank), 15. ———. 2018. “Poverty and Shared Prosperity 2018—Piecing Together the Poverty Puzzle.” (Washington, D.C.: World Bank), xi.

Kharas-McArthur-Ohno_Leave No One Behind_i-xii_1-340.indd 57 9/6/19 1:57 PM